Investigate which models are suitable for the task of dialogue classification. Generate a list of candidates for testing.

We investigate which metrics (Accuracy, Precision, etc.) are most suitable for dialogue classification. Analyse their applicability in multi-topic dialogue markup.